18 January 2016

What If a Space Shuttle Orbiter Had to Ditch? (1975)

Final approach: the inner and outer wing flaps and body flap are visible at the aft end of the descending Orbiter. Image credit: NASA
High on any pilot's list of things not to do with an aircraft is to ditch - that is, to make an emergency landing in water. Planes that nimbly slip through air become about as graceful as a brick when they touch an ocean swell. Anyone who has flubbed a dive and belly-flopped knows how painfully hard water can be.

Aircraft ditching behavior became of great interest in the United States during the Second World War, when B-24 Liberator bombers damaged by enemy fighters and flak - or simply lost and low on fuel - ended otherwise successful bombing runs over Nazi Germany by ditching in the English Channel or North Sea. Even in calm seas, the B-24 did not fare well. More often than not the plane broke apart and sank within minutes.

The Arsenal of Democracy: Workers assemble B-24 Liberator aircraft in the Consolidated Vultee plant in Fort Worth, Texas. More than 18,000 B-24s flew a wide range of missions in all Second World War operational theaters. Image credit: Wikipedia
Because of this, on 20 September 1944, four months past the beginning of the D-Day invasion of Europe, a B-24 Liberator with two brave airmen on board intentionally ditched in the calm waters of Virginia's James River. Close by in a small boat was a team of engineers from the National Advisory Committee for Aeronautics (NACA) Langley Aeronautical Laboratory in nearby Hampton, Virginia. As the 44,100-pound, four-engine, straight-wing aircraft skimmed the water's surface at 97 miles per hour with its landing gear up, instruments inside the fuselage collected data on motion and deceleration.

The B-24's belly touched water at a point just behind the wing trailing edges. The plane began to skip, its nose rising and dipping to the left; then the water seemed to grab the bomber hard. Inside, the crew felt deceleration equal to 2.6 times the force of Earth's gravity. It threw them forward against their safety harnesses. Propellers still whirling, the plane nosed down and pieces flew out of an enormous cloud of spray that momentarily hid it from view. Rescuers moved in quickly; meanwhile, both pilots climbed atop the aircraft.

The men were in good shape, but their plane, a veteran of several European bombing missions, would never fly again. Even under the relatively benign conditions of the test, its fuselage had cracked, nearly breaking in two. The crack acted as a hinge, so the plane floated, rapidly filling with water, with both its tail and its nose in the air. The right inboard motor was gone, sheared away upon contact with the water.

When the NACA engineers lowered themselves inside to recover their instruments, they were in for a shock. Almost every piece of equipment bolted to the interior of the B-24 had torn loose and been flung forward, forming a nearly impassable heap just behind the cockpit. They found their instruments, secure in water-tight containers; meanwhile, a U.S. Navy salvage boat with a crane moved in fast to hoist the plane out of the water before it joined its missing engine 30 feet down on the muddy bottom of the James River. Smaller boats collected floating pieces.

On the salvage boat's deck, the plane looked even worse than it had in the river. A large dent marked where its belly first touched the water at a descent rate of 1.8 feet per second. Though they had been reinforced for the ditching test, the bomb bay doors had been pressed inward. The bomber's thin skin was rumpled over large areas; where it wasn't creased and puckered, it was ripped.

The 20 September 1944 ditching test became a pivotal event in aerospace history. It brought home to engineers as never before the powerful forces that ditching brought to bear on aircraft. It led NACA Langley to study ditching behavior in many types of aircraft. Because ditching full-size planes was both costly and dangerous, the lab developed techniques for testing scale-model airplanes in a water trough in what became known as the Langley Impacting Structures Facility (LISF). Above all, their experiments showed that ditching was actually crashing.

Fast forward 30 years to 1975. The era of scale-model experiments was gradually drawing to a close - computer models of complex phenomena, though still crude and costly, had made their debut in aviation and other fields. When NASA had formed in 1958, Langley had become one of its research centers. With the splashdown of the Apollo-Soyuz Test Project Apollo Command Module in the Pacific Ocean in July 1975, the era of U.S. space capsules was over. The era of wings in Earth orbit was about the begin.

The Space Shuttle stack as envisioned in 1975 with major components indicated. Image credit: NASA
NASA and its contractors envisioned several plausible scenarios which might lead a Space Shuttle Orbiter to ditch. The Orbiter was a glider, so it could not try again if it missed its runway at Kennedy Space Center (KSC), Florida. If the Orbiter crew realized the problem in time, they might ditch in the Banana or Indian Rivers or in the Atlantic Ocean off Cape Canaveral.

In the event that two of the Orbiter's three Space Shuttle Main Engines (SSMEs) failed early in its ascent to space, the Orbiter would need to return to KSC; however, depending upon when the engines failed, it might not have enough altitude and energy to turn around, line up with the Shuttle runway, and stretch out its descent. In that case, it would probably fall short of the coast.

In October 1975, William Thomas, a former Langley engineer who had taken a job with Grumman Aerospace Corporation in Bethpage, New York, published results of 67 tests of a 1/20-scale model of the Space Shuttle Orbiter. The tests occurred in the LISF starting in 1974. The actual Orbiter was planned to be 37.2 meters (122 feet) long, so the tests saw a 1.86-meter (6.1-foot) fiberglass and balsa-wood Orbiter launched into a broad trough of water using a ceiling-mounted "catapult" device. The trough could simulate smooth seas or seas with swells and waves.

The model included a replaceable balsa insert designed to give some sense of the belly damage a ditching Orbiter could expect. Weights enabled Thomas to simulate either 32,000-pound or 65,000-pound cargoes in its cargo bay. With the lighter cargo, the full-scale Orbiter would weigh 85,464 kilograms (188,247 pounds); with the heavier, 103,200 kilograms (227,313 pounds). The corresponding model weights were 10.68 kilograms (23.53 pounds) and 12.9 kilograms (28.41 pounds). Small lead weights permitted tests at intermediate Orbiter weights and allowed Thomas to trim the model so that it would, for example, dip one wing as it approached the water.

The model also included adjustable flaps and landing gear which could be installed to simulate gear-down ditching or left off to simulate ditching with landing gear doors closed. Flaps and landing gear were designed to break away at scale stresses - for example, on a full-sized Orbiter, the main landing gear would fail under a load of 356,270 pounds. The 1/20-scale landing gear would break if subjected to a torque of 6.41 pounds.

In the first of the 67 tests, the model Orbiter's nose was pitched up 16°, its aft-mounted body flap - located below the SSME engine bells - was tilted down 11.7°, and its wing flaps (inner and outer) were tilted up 4°. The water in the LISF trough was calm.

The 1/20-scale Orbiter, with a simulated mass of 93,000 kilograms and a simulated speed of 53.5 meters per second (just 120 miles per hour - slow for a Shuttle landing), contacted the water with its landing gear up. It skipped, then sank deeper on the next contact. The model decelerated very rapidly - it stopped four fuselage lengths from where it first touched the water. In the full-size Orbiter, this would have amounted to about 160 meters (490 feet).  It was, Thomas commented, a "very stable run."

Beginning with Test 4, the model Orbiter was fitted with instruments for recording normal (fore-aft) and longitudinal (left-right) deceleration. These revealed that even a perfect ditching would likely harm the Orbiter crew. Apart from the instrumentation, the Test 4 Orbiter model was configured exactly like the Test 1 model. It touched calm water with its nose pitched up 12° moving at a scale speed of 72 meters per second (161 miles per hour).

Had it been a full-scale Orbiter carrying a crew, they would, after the initial skip, have been thrown forward against their straps with a force equal to 8.3 times the pull of Earth's gravity. As the Orbiter swerved to a stop, they would have felt a longitudinal jolt of 4.5 gravities. Test 4 was, it would turn out, only a little more arduous than average.

For Test 5, conditions were virtually identical to those of Test 4. Landing speed was slightly higher at 75.1 meters per second (168 miles per hour). Yet had the model been a full-scale Orbiter, it would have subjected its crew to 19.4 gravities of deceleration when it made its second contact with the water. The model's inner wing flaps broke free; this hinted that structural damage on the full-scale Orbiter was likely.

The first test of an Orbiter model with a simulated 65,000-pound payload (Test 17) saw a scale deceleration of nearly 11 gravities. In all, 14 of the 67 tests subjected the model Orbiter to greater than 10 gravities of scale deceleration.

Tests 9 and 62 were without question the most dramatic. In both, the model Orbiter stalled following release and hit the water tail first.

Other tests saw the 1/20-scale Orbiter model skip along simulated 2.1-meter (seven-foot) wave crests, plow through waves with a series of sharp jolts, dive under the water and bob to the surface, and lose both its inner and outer wing flaps. Tests with landing gear down never went well; the gear always broke away. In a full-scale Orbiter, tearing away the landing gear would likely have permitted water to enter through the damaged wheel wells.

Thomas's interpretation of the results of the 67 tests was notable: "a fairly smooth runout is expected but considerable fuselage tearing and leaking or flooding will occur." He was confident, however, that a full-scale Orbiter would remain afloat if its wings, which contained hollow spaces, remained intact. The more damage the wings suffered, the faster the Orbiter would sink.

The LISF 1/20-scale Orbiter tests were, of course, simplistic. They did not reflect the fact that Space Shuttle Orbiters were not designed to withstand the deceleration loads most of the model tests indicated. The average ditching deceleration ranged between five and eight gravities; this would have been sufficient to cause significant structural damage. In other words, in almost all cases, the Orbiter would have snapped apart.

Moments before the end: the plume from the Solid Rocket Booster leak that doomed Challenger is clearly visible. Image credit: NASA
A little more than a decade later (28 January 1986), the Challenger accident made obvious the Space Shuttle Orbiter's fracture lines. The Orbiter Challenger did not explode; rather, the brown External Tank (ET) upon which it rode and from which it drew liquid hydrogen/liquid oxygen propellants for its SSMEs was destroyed by a malfunctioning Solid Rocket Booster (severe wind shear might also have played a role).

The fuel and oxidizer the ET contained came together and ignited, producing an explosion, but it was aerodynamic forces that tore the Orbiter apart. Basically, as its ET disintegrated, Challenger's nose stopped pointing the right way. This subjected it to drag and deceleration.

The crew compartment broke free, trailing behind it a comet's tail of cables. Some of the crew inside remained conscious long enough to take prescribed - though futile - emergency measures. The compartment remained mostly intact until it hit the water about two minutes later.

The Payload Bay disintegrated, but Challenger's main payload, a Tracking and Data Relay Satellite with an attached Inertial Upper Stage, remained more or less intact as it flew free of the fireball and fell toward the Atlantic. Challenger's wings separated, partly disintegrating. Each, however, remained recognizable in images and video captured from the ground. The aft compartment containing the SSMEs also emerged from the fireball mostly intact.

After Challenger, the Space Shuttle Program came under intense scrutiny. The Report to the President by the Presidential Commission on the Space Shuttle Challenger Accident cited Thomas's 1975 report when it stated that the probability was high that a Shuttle Orbiter would break up and sink after a ditching. Even if the Orbiter remained intact, the Commission report continued, cargoes mounted in the Payload Bay would break loose from their supports, slide forward, and smash into the back of the crew cabin. Though Thomas's report did not in fact say these things, they were logical conclusions based on the results of the 67 1/20-scale model Orbiter test runs.


YouTube - "Ditching of a B-24 Airplane into the James River" - https://youtu.be/WjadMxpXprk (uploaded by Jeff Quitney - accessed 16 January 2016)

YouTube - "Space Shuttle Orbiter Ditching Investigation of a 1/20-Scale Model" - https://www.youtube.com/watch?v=dlG-HcZIDl0 (uploaded by Jeff Quitney - accessed 16 January 2016)

Ditching Investigations of Dynamic Models and Effects of Design Parameters on Ditching Characteristics, Report 1347, L. Fisher and E. Hoffman, Langley Aeronautical Laboratory, National Advisory Committee for Aeronautics, 1958

Ditching Investigation of a 1/20-Scale Model of the Space Shuttle Orbiter, NASA Contractor Report 2593, W. Thomas, NASA, October 1975

Report to the President by the Presidential Commission on the Space Shuttle Challenger Accident, Presidential Commission on the Space Shuttle Challenger Accident, Volume I, pp. 182-183, June 1986

"A Plane Crash in 1944 Is Saving Lives Today," Peter Frost, Daily Press, 22 February 2009 - http://articles.dailypress.com/2009-02-22/news/0902210116_1_b-24-successful-emergency-landing-hudson-river (accessed 16 January 2016)

More Information

What If an Apollo Saturn Rocket Exploded on the Launch Pad? (1965)

What If a Lunar Module Ran Low on Fuel and Aborted Its Landing? (1966)

Where to Launch and Land the Space Shuttle? (1971-1972)

What Shuttle Should Have Been: NASA's October 1977 Space Shuttle Flight Manifest

One Space Shuttle, Two Cargo Volumes: Martin Marietta's Aft Cargo Carrier (1982)


  1. The ocean. Not recommended. I live on the Atlantic. More precisely near the "Graveyard of the Atlantic", Cape Hatteras. Florida is near the Gulf Stream. Also quite unsettled. Normally 4-6 foot swells on a calm day.

    Which reminds me. If Elon ever gets his rocket to land on a pitching barge, that will be a feat.

    If a shuttle had to ditch, lord help them. But we're done with that now. I liked the discussion of Challenger. Had the Challenger "dismounted" before the center tank exploded, we might still have some 7 jostled, but breathing astronauts now. That big flat bottom of the orbiter would have made quite the surfboard.

    1. Ben:

      The trouble is, the Orbiter was structurally fragile. It had to do what it was meant to do, stay pointed as it was meant to be pointed, touch down as it was meant to touch down, or in almost all cases the vehicle was lost and the crew killed. We didn't spend enough to give it adequate margins.

      The Presidential Commission noted that NASA had added a "fast separation" option to the Orbiter software immediately after Challenger. However, it noted also that in all computer model testsof the new abort mode, the Orbiter hung up on its aft ET attach points and pitched backwards. Basically, it began an end-over-end tumble - but the Orbiter's frailty meant that it wouldn't complete even one complete somersault before it broke apart.


    2. Interesting. Makes alot of sense about the earlier designs with a smaller orbiter for personnel only mounted on an expendable rocket. Any explosion would have been from behind and escape technology of the day would have been a follow-on from the escape towers of Apollo, possibly.

      An alternate cargo configuration would have been for larger items, and I think you covered that in part with an earlier post on the Saturn IV-B.

      Its a shame the Air Force didn't get to finish MOL and Dyna-Soar. That technology seems a good bridge between Apollo and Shuttle. MOL would have been analogous the Salyuts of the day, I suspect.

      Ah Nixon. One wonders what a Johnson or Humphrey space program would have been from 1968-76.

  2. Remarkably, there is a film of the B-24 ditching on youtube: https://www.youtube.com/watch?v=ggraQgZGe4Y

    1. Stuart:

      You'll note that I have a link to a YouTube video showing the 20 September 1944 B-24 experiment at the top of my sources. Just below that is a link to video of some of the 1/20-scale Orbiter test runs. The 2009 newspaper piece at the bottom of my sources contains reminiscences of one of the Langley participants in the 1944 test.



I like hearing from my readers. No rules except the obvious ones - please keep it civil and on topic.